vrend, shader: emit all the generic outputs the next shader expects

Pass a mask of the non-array generic inputs that the next shader expects
is passed to the shader that is currently converted. If, after emitting
all generic outputs, some are missing, then these are also generated.

Limitations: This doesn't take care of input arrays that may not be emitted
as outputs, but since this problem seems to be only related to IO variables
that are implicitely declared this is not a problem.

Fixes piglit: glsl-routing

v2: rebase and update names to new naming

v3: Declare TCS outputs as arrays as required

Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Tested-by: Elie Tournier <elie.tournier@collabora.com> (v1)
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

src/vrend_renderer.c

@@ -3043,6 +3043,7 @@ static inline void vrend_fill_shader_key(struct vrend_context *ctx,
    if (next_type != -1 && ctx->sub->shaders[next_type]) {
       key->num_indirect_generic_outputs = ctx->sub->shaders[next_type]->sinfo.num_indirect_generic_inputs;
       key->num_indirect_patch_outputs = ctx->sub->shaders[next_type]->sinfo.num_indirect_patch_inputs;
+      key->generic_outputs_expected_mask = ctx->sub->shaders[next_type]->sinfo.generic_inputs_emitted_mask;
    }
 }
 

src/vrend_shader.c

@@ -162,6 +162,10 @@ struct dump_ctx {
    struct vrend_io_range generic_output_range;
    struct vrend_io_range patch_output_range;
 
+   uint32_t generic_outputs_expected_mask;
+   uint32_t generic_inputs_emitted_mask;
+   uint32_t generic_outputs_emitted_mask;
+
    uint32_t num_temp_ranges;
    struct vrend_temp_range *temp_ranges;
 
@@ -5596,6 +5600,14 @@ emit_ios_generics(struct dump_ctx *ctx, enum io_type iot,  const char *prefix,
                 t,
                 io->glsl_name,
                 postfix);
+
+      if (io->name == TGSI_SEMANTIC_GENERIC) {
+         if (iot == io_in)
+            ctx->generic_inputs_emitted_mask |= 1 << io->sid;
+         else
+            ctx->generic_outputs_emitted_mask |= 1 << io->sid;
+      }
+
    } else {
       if (prefer_generic_io_block(ctx, iot)) {
          const char *stage_prefix = iot == io_in ? get_stage_input_name_prefix(ctx, ctx->prog_type):
@@ -6116,6 +6128,18 @@ static void emit_ios(struct dump_ctx *ctx)
       return;
    }
 
+   if (ctx->generic_outputs_expected_mask &&
+       (ctx->generic_outputs_expected_mask != ctx->generic_outputs_emitted_mask)) {
+      for (int i = 0; i < 31; ++i) {
+         uint32_t mask = 1 << i;
+         bool expecting = ctx->generic_outputs_expected_mask & mask;
+         if (expecting & !(ctx->generic_outputs_emitted_mask & mask))
+            emit_hdrf(ctx, "                              out vec4 %s_g%dA0_f%s;\n",
+                      get_stage_output_name_prefix(ctx->prog_type), i,
+                      ctx->prog_type == TGSI_PROCESSOR_TESS_CTRL ? "[]" : "");
+      }
+   }
+
    emit_ios_streamout(ctx);
    emit_ios_common(ctx);
 
@@ -6277,7 +6301,7 @@ static void fill_sinfo(struct dump_ctx *ctx, struct vrend_shader_info *sinfo)
       free(sinfo->image_arrays);
    sinfo->image_arrays = ctx->image_arrays;
    sinfo->num_image_arrays = ctx->num_image_arrays;
-
+   sinfo->generic_inputs_emitted_mask = ctx->generic_inputs_emitted_mask;
 }
 
 static bool allocate_strbuffers(struct dump_ctx* ctx)
@@ -6349,6 +6373,7 @@ bool vrend_convert_shader(struct vrend_context *rctx,
    ctx.has_sample_input = false;
    ctx.req_local_mem = req_local_mem;
    ctx.guest_sent_io_arrays = key->guest_sent_io_arrays;
+   ctx.generic_outputs_expected_mask = key->generic_outputs_expected_mask;
 
    tgsi_scan_shader(tokens, &ctx.info);
    /* if we are in core profile mode we should use GLSL 1.40 */

src/vrend_shader.h

@@ -67,6 +67,7 @@ struct vrend_shader_info {
    uint8_t num_indirect_patch_outputs;
    uint8_t num_indirect_generic_inputs;
    uint8_t num_indirect_patch_inputs;
+   uint32_t generic_inputs_emitted_mask;
    int num_ucp;
    int glsl_ver;
    bool has_pervertex_out;
@@ -117,6 +118,7 @@ struct vrend_shader_key {
    uint8_t num_indirect_patch_outputs;
    uint8_t num_indirect_generic_inputs;
    uint8_t num_indirect_patch_inputs;
+   uint32_t generic_outputs_expected_mask;
 };
 
 struct vrend_shader_cfg {